Abstract
The human perceptual system extracts socially-relevant information from biological motion, and observers can perceive emotions, gender and attractiveness even from point-light displays. Attractiveness is an evolutionary important social trait, which has been studied extensively with static pictures of faces. Previous work shows that individuals acrossdifferent cultures tend to agree on which faces are attractive, and it has been hypothesized thatevolutionary pressure explains the preference for average, symmetrical, and feminized faces. For female faces perceived attractiveness and femininity are closely related: Feminized faces are consistently ratedas more attractive, both by male and female raters. This raises the question how gender and attractiveness are related for dynamic stimuli, such as gait patterns.
METHODS: We investigated the relationship between sexual dimorphism and perceived attractiveness in gait patterns by presentation of movies showing volumetric puppets with standardized body shape, which were animated using motion capture data. Perceived attractiveness and gender for the individual movies were rated by 10 male and 10 female observers on 7-point likert scales.
RESULTS: Gender and Attractiveness were consistently rated and the correlation between these ratings was very small and non-significant (r = 0.038, p [[gt]] 0.5). Inspired by this result, exploiting motion morphing methods, we constructed a stimulus set that permits an independent variation of the attractiveness and gender of walkers. Controlling for the influence of gender, attractive walks are characterized by high symmetry and regularity, while unattractive walks showed strong asymmetries and often resembled pathological gaits.
CONCLUSION: We failed to find the relationship between sexual dimorphism and attractiveness in dynamic patterns that was previously reported for faces. Instead, attractiveness and gender seem to form separate dimensions in the encoding of social traits in dynamic gait patterns. This implies potentially different constraints for the communication of social signals by faces and body motion.
Funded by the EC FP6 project COBOL, and the DFG.